The goal of this project is to identify the genes (and their respective sequence variants) that play a role in conferring susceptibility to obesity in humans. Knowledge of the nature of these genes, and of the biological mechanism(s) of their actions, would contribute enormously to diagnosis, treatment and prevention of obesity and its allied disorders such as diabetes, cardiovascular disease and certain cancers. 1.) Novel candidate genes influencing adiposity and energy homeostasis will be identified by ongoing surveillance of the literature and by studying large-scale differential gene expression in rodent and human models of altered energy expenditure using cDNA expression microarrays. 2.) Highly sensitive techniques for detecting DNA variants in both coding and predicted regulatory sequences for these candidate genes will be applied to a collection of 2,000 subjects selected for extreme and/or early onset obesity. Allele frequencies of these DNA variants will be determined in four major ethnic groups. Based upon these analyses, phenotypically well-characterized subjects of suitable ethnicity will be selected for association studies from over 88,000 individuals available through 28 collaborating investigators around the world. Statistical analysis will include joint tests of linkage and association using transmission disequilibrium testing, new methods of genomic control, and haplotype analysis. We are also developing new methods to analyze pooled data from multiple populations and to model multiple loci with epistatic interactions. 3.) For variants with positive association results, assessments of possible functional significance will be made using computer modeling and predictions of function, as well as in vitro and in vivo (transgenic) functional assays. In cases in which the alleles associated with increased adiposity have no apparent functional consequence, the entire intronic, exonic, and 5' and 3' regions of the gene will be sequenced in 20 (15 obese; 5 lean) selected subjects to find non-coding variants that account for the phenotypic variation. Coding and non coding sequence variations will be examined for functional effects using in vitro expression techniques and transgenic mice with graded/inducible expression and human allele substitution. Finally, technologies will be developed to rapidly and inexpensively genotype large numbers of subjects for a panel of obesity candidate gene SNPs. Twenty five new genes will be studied over a 5 year period. A great strength of the proposed approach is access to large numbers of phenotypically well characterized subjects of varying ethnicity. These numbers will likely be required to detect sequence variants with specific effects on particular subphenotypes, small phenotypic effects, ethnically-specifc genotypes or epistatic interactions among genes.